Le prochain séminaire qui aura lieu le lundi 6 juillet 2026 à 14h en B014 (et pas visio) sera donné par Mosab Bazargani, lecturer en Data Science et IA à l’université de Bangor, UK.
Titre:
Robust augmentation and learning for accurate discrimination of PFAS containing material in high noise Raman Spectroscopy
Abstract:
The detection and classification of per- and polyfluoroalkyl substances (PFAS) are becoming increasingly important for supporting sustainable manufacturing and regulatory compliance across the aerospace industry. This talk presents a deep learning-based approach for accurate multi-class classification in scan-and-classify industrial applications using Raman spectroscopy. At the core of the framework is a data augmentation protocol that enhances the characteristic features of known PFAS spectroscopic scans, improving robustness to noise and increasing the diversity of the training data. Experimental results on real Raman spectra collected from inorganic aerospace components containing PFAS achieved classification accuracies exceeding 95%, with sub-second inference times suitable for real-time deployment. The compact model architecture is well suited for deployment on remote or resource-constrained devices, while its efficient training enables frequent server-side retraining as new data become available. The talk will also discuss the practical challenges of deploying AI models in industrial environments and highlight how this research, developed in collaboration with the Sustainability Team at Airbus Defence and Space, supports the transition towards more sustainable aerospace ymanufacturing.
Bio:
Dr Mosab Bazargani is a Lecturer in Data Science and AI at Bangor University, UK. He received his PhD from Queen Mary University of London, where his research focused on the intersection of Artificial Intelligence and Operational Research, developing optimisation methods for large-scale industrial problems in scheduling, routing, logistics, and sustainability.
Before joining academia full-time, he worked as a Data Scientist at Tesco PLC, where he was part of the team developing optimisation solutions for Tesco’s secondary transportation network, contributing to systems that deliver multi-million-pound annual savings. He later worked as a Data Analytics Contractor at Marsh McLennan. Since joining Bangor University, he has led and contributed to AI and optimisation projects with industrial partners including Airbus, KLM, Mace (Sizewell C), and Rolls-Royce, as well as collaborating with the Welsh Government on AI-driven initiatives supporting the transition to Net Zero 2050. He was also one of three members of the Bangor University team that helped secure the UK Government’s AI Growth Zone initiative for North Wales, supporting the development of a major AI Data Centre in the region.